49222e2c9fa43dc101efb72ab71c94fe94e41089,tensorforce/tests/test_dqfd_agent.py,TestDQFDAgent,test_dqfd_agent,#TestDQFDAgent#,34

Before Change

        // First: add to demo memory
        for n in xrange(50):
            action = agent.get_action(state=state)
            if action == 0:
                state = (1, 0)
                reward = 0.0
                terminal = False
            else:
                state = (0, 1)
                reward = 1.0
                terminal = False
            agent.add_demo_observation(state=state, action=action, reward=reward, terminal=terminal)

        // Pre-train from demo data
        agent.pre_train(10000)

After Change

        agent = DQFDAgent(config=config, network_builder=network_builder)

        // First: generate some data to add to demo memory
        state = environment.reset()
        agent.reset()

        for n in xrange(50):
            action = agent.act(state=state)

Instances